ADDRESSES. about ten grams in a platinum dish over a Bunsen burner for five hours. The second determinations were made b y heating about ten grams in a nickel crucible for fifteen minutes with a blast lamp. The above data are presented not only t o show the wriations due to sampling and assaying, but also t o give the retail pharmacist some reason for the lack of uniformity of his tincture of asafetida. Perhaps the present method of valuation is far from ideal, but certainly some definite standard for the amount of soluble constituents in the tincture should be insisted upon. Acknowledgment are due Mr. H. M. Sechler for considerable of the data here presented.
423
He had mapped out extensive lines of investigation and had initiated several different researches of a most difficult nature. Unfortunately, his withdrawal from the bureau has hindered the prosecution of some of this work for lack of a directing spirit with full comprehension of the subject in its different bearings. At the same time it was felt, in view of the obscurity in which some of the questions were involved and the length of time needed for their study and the uncertainty of a successful issue, that our chief attention should be centered for the present on other lines of work where success was practically assured and the immediate value to the community was plainly evident. The most important of these problems was a great extension of one already initiated under Dr. Noyes, namely, the providing of carefully analyzed materials by which chemists can check the accuracy of their analytical methods and employers can control the work of their employees. The usefulness of these materials is not confined to those employed in ANALYTICAL DEPARTMENT, the industries but extends equally to educational institutions. SMITH,KLINE& FRENCH Co To you, being chemists, it is hardly necessary to offer any . ~ arguments in support of the use of such standardized materials. The subject is treated briefly, but with sufficient fullness, in a ADDRESSES. publication which has been issued only very recently a s a circular of information regarding the samples that have been CHEMISTRY IN THE BUREAU OF STANDARDS. provided thus far by the bureau or are in immediate preparaBy W. F. HILLBBRAND, Chief Chemist. tion, some thirty-seven in all. The number will be very largely Received September 7 , 1910. I have felt impelled to speak on the subject “Chemistry in added to, for there is a n ever increasing demand in this directhe Bureau of Standards,” for the reason, mainly, that since tion, and I feel that the bureau can do no greater service to chemists of the industries a t present than by catering to this my connection with t h a t bureau little published evidence of demand. Not only will benefit result from the use of the samchemical activity has been offered. The impression might ples themselves but also from the experience that the bureau arise and spread that proper advantage is not taken of the chemists will gain, and have already gained, in their close study opportunities provided. It is my wish to forestall such an opinion. That the public closely interested in chemistry has of the methods employed in analyzing them. To this feature I will revert later. a right to be informed on the subject goes without saying, The list of samples in stock comprises three pig irons, 5 Bessefor the bureau is a federal institution where work is expected mer steels, 5 basic open-hearth steels, 6 acid open-hearth steels, t o be directed chiefly toward providing the public with authentic and I vanadium steel, to which will be added, as soon as analinformation on a variety of subjects. This information must ysis can be made, a nickel steel, a chrome-vanadium steel, and be of value for the most part to the industries, otherwise the a chrome-tungsten steel. Of ores there are a n argillaceous bureau fails in its chief function. From its inception the bureau has endeavored to meet the most insistent demands upon its limestone, suitable for the manufacture of Portland cement, reservoir of physical knowledge, and at an increasing rate as a zinc ore, and 3 iron ores from the Lake Superior region. It is impossible to find any single ore from this region that can facilities were given it for so doing. That i t has been in the advantageously be used a s a standard for all. the constituents main eminently successful in this respect, is sufficiently attested by the support it has a t all times received from Congress and b y * that are ordinarily determined in a n iron ore. Hence, the the marked growth it has experienced and is experiencing, a multiplication, one being used for iron, phosphorus and silica, growth that would be impossible without strong backing, that the second for alumina, lime and magnesia, and the third for manganese. A titaniferous iron ore from New Jersey is in process is, the approval of a n influential section of the community of analysis, and it is proposed to provide later one typical ore both industrial and educational, manifested through its reprefrom the Alabama region. Possibly typical ores of lead and sentatives in the national capitol. The growth of the chemical side of the bureau has been slow, zinc from the Rocky Mountain region will follow. I t is to be said here, however, that the selection of ores that but it is now increasing rapidly a t a rate which bids fair to call for a sFecial building within a cwple of years. At the start can be used advantageously as standards for analysis offers it was not clear, apparently, along what lines the. chemical far greater difficulty than that of alloys, for the reason that they are almost always so subject to changes in moisture conwork of the bureau might best develop and I mas myself in tent from day to day or at different altitudes, as to detract doubt for some time. The viewpoint of my predecessor, Dr. very much from their value. If their sensitiveness to hygroNoyes, that research without direct practical bearing should metric variations is at all pronounced it is almost hopeless to not be elbowed out of view by the demands of the industries, is one that should need no defense before a gathering of this expect that all chemists will be able to reduce them to a fixed character. It is a view that is held by the director of the water content, no matter how precisely formulated may be the bureau a s well as by myself. Xevertheless, during the interval directions for drying. The moisture content may vary somewhat without detriment so far as the minor constituents are between the withdrawal of Dr. Noyes from the bureau and my concerned, but not so for those present in large amounts own transfer to it, the work of the small chemical force was wholly along practical lines. The demands of one and anWe have also a pure sugar for use 0s a standard in caloriother of the government bureaus and departments for help in metric and saccharimetric work, and studies are in progress the preparation of specifications for materials in which they looking to the applicability of other standards for the measurewere vitally interested had become insistent. Dr. Stokes, ment of heat of combustion, such as benzoic, salicylic, phthalic the associate chemist in charge, was deeply impressed with and hippuric acids, naphthalene, anthracene and camphor. the importance of the numerous problems presented and de- This work is done in co-operation with the heat division of voted himself with ardor and great ability to their solution. the bureau.
424
T H E JOURNAL OF INDC'STRIAL A N D ENGI:VEERIiVG CHELVIISTRY.
The American Brass Foundrymen's Association has enlisted our aid in the preparation of standard brasses and bronzes, and a beginning has been made with two brasses, which are in preparation We have undertaken to aid the fertilizer interests by carefully analyzing one or more typical phosphate rocks and by testing a t the same time the relative values of the different methods in common use, especially with respect to the influence on them of such interfering constituents as are common to materials of this class-pyrite, fluorine and organic matter. I could easily devote an hour to the problems encountered in the selection and preparation of thexsamples themselves and several more to those connected with their analysis An enormous volume of correspondence on these points has accumulated We endeavor to keep in close touch with prominent manufacturers, and especially a i t h prominent chemists, soliciting their advice and criticism Almost without exception our efforts have met with hearty approval and support I t may be well to introduce here some remarks relative to the plan we have hitherto followed in obtaining the composition of our samples. I t is a plan of co-operation, in which leading works chemists and commercial chemists are invited to participate The almost inevitable result is failure of satisfactory agreement in the reported results Ordinarily n e have given opportunity to those whose results differed markedly from the mean to revise their work without indicating to them the direction or magnitude of the deviation This course has often resulted in much loss of time and necessitated much correspondence For certain cases, and eventually perhaps for all, the bureau may decide to depend solely on its own chemists to establish the composition of a given material, but for some time to come this plan will be employed to a limited extent only The bureau must first afford to its analysts sufficient time to become masters in their particular fields to an extent that will command general confidence Until this condition is fulfilled i t is far better policy to share the responsibility with experts outside of the bureau. The certificates which the bureau issues with its analyzed samples represent, therefore, the best efforts of a considerable number of analysts who are supposed to be especially expert in the analysis of the given material, and the mean undoubtedly gives in most cases a close approximation to the correct composition The bureau does not, however, undertake to vouch for their correctness The results by different analysts and by different methods are reported in such a way that the person using them can see at a glance what variations are to be expected in commercial work done by supposed experts and under conditions that are presumably the best that are to be encountered in commercial laboratories In certain crtses, such as pig iron and the ordinary steels, the methods commonly used and the precautions observed are of the highest known order. The commercial results on the average are, therefore, to be accorded fully as great weight as those obtained in the bureau I t is not always so with other materials, when the character of the reagents or the apparatus play a more important part, and when the complexity of composition, as in ores, introduces special difficulties in the x a y of effecting complete separation of the constituents and their accurate determination I n special cases the bureau regards its own values as closer to the truth than the mean of any number of outside determinations. This is true, for instance, of the alumina and magnesia in one of the Lake Superior iron ores and even in the value for iron itself in another of them. One without experience can form little idea of the dificulties and delays that have been encountered, from the initial selection of one of our samples through the machining, grinding, mixing, analyzing, and assembling of data to its final issuing. It took us nearly a year to get the bars to replace one of our
Oct., 1910
steels of which the supply gave out. The steel company that undertook to fill the order had to make five castings before + product of the right composition was obtained. Once delivered to the bureau, it takes from two to three weeks to reduce the sample of steel or iron to fine chips These sometimes have to be ground finer in a special machine and in any case must be most thoroughly mixed. The successful mixing of 300 or 400 pounds of such a material requires a special mechanical mixer and no one of those on the market is entirely satisfactory The one we have used we modified materially and are now having one made that works on a different principle entirely. \%'e have experimented somewhat with a view to securing homogeneous samples of relatively low melting alloys and metals in a state of sufficient division, without the use of a lathe, and have some hopes of eventual success. Much time has to be expended by us in analyzing these materials, particularly the ores, brasses and special steels, for which the commercial methods are either more or less inadequate or have not yet received the extended study and testing that has been given to those for irons and plain steels I t is to be borne in mind that we must not be content with a single determination or with a few determinations by a single method, but that our problem is to ascertain just as closely as is possible the true composition of these complex materials. Literally months of time may thus be consumed in the study of a single sample of a new kind. The knowledge thus gained, however, shortens by far the time required for later analyses of similar or related samples. In acquiring this knowledge and in testing the various commercial methods against each other, we have already made useful observations that will be of value to others as well as to ourselves. e& '\ shall not hasten to publish our observations, but wait until we haye had full opportunity to verify them repeatedly. Much information on certain of our own methods and in less degree on those used by others is to be found in our recently issued circular- 'Methods of Analysis for Iron and Manganese Ores." In ac!dition to the above, me have taken the preliminary steps looking toward the issuance by the bureau of a sodium oxalate for the standardization of solutions to be used not only in oxidimetry, but in acidimetry as well. Our reasons for making choice of sodium oxalate I need hardly go into now, save to say that for oxidimetry it seems to be the most satisfactory material that can be had in considerable quantity. Il'hen pure (and the different makes on the market are not to be taken a t their face value, but must be tested by each buyer) the article is stable and eminently satisfactory as a standard. The use of so-called piano wire as a basis for s andardizing permanganate and bichromate solutions cannot be too strongly condemned. Electrolytic iron is difficult to prepare and it too is not as pure as could be wished. The chief difficulty before us is to secure from manufacturers an article of sufficient purity in the large amounts called for If found, it will be tested by us and issued under our seal Should success attend our efforts in this case, we may go farther. Having thus dwelt at some length on that feature of our work that is likely to appeal most strongly to chemists as a whole, I will pass over others more briefly Various bureaus of the government have called on the Standards Bureau for assistance. These requests range from simple requests for tests of materials to those that involve more or less extended research. Six months were required for a research on pyroxylin plastics with reference to their acceptance or rejection as freight on American passenger vessels. The Treasury Department wants information to enable it to draw specifications for its various record inks and to improve their quality. The Government Printing Office needs similar information in regard to its papers and inks. I n fact the Standards Bureau has for
ADDRESSES. some years done the greater part of the physical and chemical testing of papers for that great printing establishment, and frequently analyzes the metals and alloys that enter into the composition of type metals or find direct use in printing. This work requires not only the time of one chemist for‘ the routine work but of another for research connected with paper. The following list embodies tests that were made by us in one year: Papers, writing and printing inks, mucilage, lubricating oils, fuel oil, linseed oil, turpentine and white lead, paints, gelatine compound, boiler compounds, soaps, rubber hose and valves, silks, brines, electroplating solutions and scrapings from anode plates, flooring compositions, silicate brick, glasses, gypsum, boiler tubes, steels (other than those in our list of analyzed samples), graphite, tinfoil, solder, monotype metal, antimonial lead, jute bagging, rasped pine sawdust cover for paper rolls, condenser tubes and water, guide rails. Many of these called for far more than routine testing. In consequence of recent legislative action by Congress, it is probable t h a t a vast deal of work in connection with the testing of structural materials will devolve upon the bureau. I t s proper attention will require large additions to the chemical force and necessitate the erection in the near future of a building specially devoted to chemistry. A s is generally known, the units in terms of which practically all electrical measurements are expressed, are the International ‘Units defined in terms of concrete standards, and not the absolute units of the C. G. S. system. The latter have been displaced, for the present a t least, on account of the higher accuracy attainable with the former by the aid of carefully drawn specifications. To secure the highest attainable accuracy, the chemist must be called upon to study the methods of purification or preparation of the materials employed, to develop special methods of analysis, qualitative a s well a s quantitative, and incidentally to investigate various problems such a s arise in connection w th eve y research. The London International Electrical Conference of 1908 adopted the international ohm and amFere, defined in term of the mercury column and the voltameter resFectively, a the two fundamental units. The IVeston normal cell, which must, however, he employed to fix the results of voltameter work, was adopted a s a secondapy unit. Owing to the difficulty of reaching agreement in specifications, this question and others were assigned to a n International Scientific Committee of 1 5 under the auspices of which a meeting of representatives of the National laboratories of Germany, England and France was recently held a t the Bureau of Standards. The object mas to compare the results obtained with the types of voltameters used in the various laboratories and to compare standard cells representing the procedures adopted in those institutions a s well a s to investigate sources of variation in the results, in order to pave the way for the adoption of uniform specifications guaranteeing the highest accuracy of reproduction. The results, though incomplete, are to be regarded a s most satisfactory, but cannot be discussed here. The services of a chemist were needed during all the work on the voltameter and for many months in advance of the coming of the foreign delegates. To his skill was due in large measure the unraveling of some most troublesome and obscure points in the action of different types of coulometers and of the solutions employed in them The main chemical questions that have a bearing on the results of voltameter measurements are the following. I The preparation of pure silver nitrate. 2 Methods of testing the purity of silver nitrate 3 The influence of inipurities in the electrolyte.
425
4. The contamination of the electrolyte by the septum employed or by the electrolysis effected. The chemical side of the reproducibility a n d , constancy of the Weston cell a s a standard of electromotive force has been under consideration at the Bureau of Standards for some time with the result that an accuracy of reproduction of two or three parts in IOO,OOO is now attainable. (Parenthetically i t may be said that a like degree of accuracy is striven for in the reproducibility of conditions and results by the voltameter, that is, of at least two or three parts in 100,ooo a s a mean of several determinations.) The work with the Ii’eston cell has embraced the investigation of equilibrium conditions in the cell, the methods of washing and the hydrolysis of the depolarizer, mercurous sulphate, the preparation and purification of the materials mercury, cadmium, mercurous sulphate and cadmium sulphate, and the effect of added impurities. I t is hoped that the ultimate result of the international comparison of cells and volta,meters and the exchange of the various materials will result in the’ adoption of uniform specifications for these materials. The bureau has made the proposal that when this shall have been effected the various national laboratories shall undertake to furnish such materials ready for use to investigators and others. The need of a new and authoritative table of densities of alcohol-water solutions has long been apparent. Fundamental for this is an exact determination of the density of absolute alcohol. The data heretofore chiefly relied on are those of Mendeleef and the Squibbs, but these differ, hence the desirability of a new determination, which necessitates the prior preparation of pure alcohol, a most difficult problem. How this has been done is described in a paper by Mr. E. C. McKelvy, which I shall have the pleasure of presenting in abbreviated form a t this meeting. An investigation is in progress having in’ view the unification and simplification of methods of gas testing in this country. An important point in connection with this relates to the permissible limit for hydrogen sulphide in illuminating gas, and the accurate and quick determination of the amount present. In co-operation with the proper physical division a chemist is a t work on this problem. Another has worked in co-operation with the heat division of the bureau in the preparation of the pure materials for calorimetric comparisons, of which I have already made mention. This latter research is closely related to another, namely, that of determining the heats of combustion of methane and ethane, ethylene and acetylene, carbon monoxide and hydrogen, which are needed a s fundamental constants for a study of the calorimetry of illuminating gases. The above-named substances must be prepared in a state of great purity and for their combustion large quantities of very pure oxygen are needed. One chemist in co-operation with our very efficient glassblower has devised a generator for oxygen and hydrogen consisting of nine cells. The electrodes are of wrought iron and are separated by a glass bell-jar. The electrolyte is caustic soda of 2 0 per cent. strength. The cells are electrolytically in series and the current used about 2 0 amperes. With this current the yield of oxygen is about 47 liters per hour. The gas is passed through a platinized quartz tube two meters long, then through two towers filled with solid potassium hydroxide, then through a soda-lime tower, and finally through a calcium chloride tower. The gas is then compressed into cylinders a t IOO atmospheres pressure by a special mercury piston pump. I t is expected that the bureau will, in future, draw- its supply of oxygen and hydrogen from this source. A comprehensive co-operative study of underground electrolysis and corrosion is to he started this year and for this the services of two chemists will probably be needed. Finally, a long investigation, still unfinished, is the deter-
426
T H E JOURNAL OF INDUSTRIAL A N D ENGINEERING CHEMISTRY.
mination of the atomic weight ratio between bromine and hydrogen. This research forms a logical extension of that already done at the bureau by Drs. W. A. Noyes and H. C. P. Weber on the ratio between chlorine and hydrogen, a work that was crowned some time ago by the award of the Nichols medal. The force of chemists now employed numbered 13 on June 3oth, which number will be increased from 4 to 6 this summer, and probably by a number of others. A t present the quarters are uncomfortably crowded, but before the summer is over additional room will be provided to accommodate 17 or 18in all. There is a weekly journal meeting for the chemists, also certain courses of lectures and laboratory work to aid those who desire to take advanced university degrees. This latter work is done out of office hours and is accepted by some universities. In the foregoing I have endeavored to outline in the briefest possible manner the character of the chemical work done at the bureau. You will see that it is most varied and that administration and correspondence must necessarily make such large demands on my own time as to leave little for direct experimentation. The most that I can hope to do must be a t odd moments, but i t is my hppe and wish to be able to exercise direct supervision over the more exact analytical work, so far as this falls in line with my personal acquaintance. The bureau will, before long I hope, do its full share in the way of publication to forward the science of chemistry both practically and theoretically. But the bureau is yet young and so are most of its chemists. Hence actiye publication from the start is not to be expected, for unusual care must be taken lest work of doubtful value shall emanate from an institution which should be what its name implies, a Bureau of Standards. I t is too much to hope that this will never happen, but my aim will be to keep it at a minimum.
NOTES AND CORRESPONDENCE.. NOTE ON T H E VOLATILITY OF COCAIN. When cocain residues .?ere dried in a covered dish heated over the steam bath or at 100’ C. in a drying oven, i t was noted that a solid film soon collected on the cover. This sublimate was frost-like in appearance, insoluble in water, and on further examination was found to consist of cocain. The fact that this body is volatile a t 100’is importan; in connection with analytical work, if the analyst expects to heat a residue after the solvent has been evaporated, and a caution to this effect would not be out of place in the pharmacopeia1 directions for assaying coca leavs. Two cocain residues were tested by heating firsr for seven hours over a steam bath and then for seven hours in an oven a t 100’C., the weights being 0.0131gram and 0.0121 gram, the dish in the case of the first sample being open to the air and the second covered with a watch glass. A t the end of the first stage of the experiment they weighed respectively 0.0108 gram and 0.0104gram and at the end of the second stage 0.0083 gram and 0.0089 gram. There seemed to be practically no difference in the proportional loss whether the dish were covered or not. A cocain residue heated several hours in a drying oven at temperatures of 60°, 80’ and 90° suffered no loss in weight, and when heated to 98’ there was only a slight loss though there was considerable sublimation upward on the walls of the dish. I t was evident from the observations that the last portions of the solvent should be evaporated as rapidly as possible from a cocain residue, using a blast; and that the residual material should be dried cautiously, not over 90° and for safety should be brought to a constant weight by drying over sulphkric acid. H. C. FULLER.
Oct., 1910
NOTE ON DR. BACKE’S INVESTIGATION OF MALTOL AND ISOM ALTOL. Edztor Journal of Industrial and Engineering Chemistry: Since the, publication of my short paper on “ A Source of Error in the Examination of Foods for Salicylic Acid,” which was sent you September, 1909,and appeared in this Journal for January, 1910,i t has come to my notice that a communication relating to the same source of error was published by Dr. Arnold Backe in Annales des falsifications for November, 1909,and Dr. Backe has recently informed me by letter that the experiments upon which his communication was based were made in July of last year. Since my own experiments were made the month following, I beg space to acknowledge the priority of Dr. Backe’s work on this point over my own, and also to call attention to his more recent papers,’ in which he discusses the isolation and properties of the substances to which the interfering reaction is due. H. C. SHERMAN.
ACCURACY IN SAMPLING. EdLtor Jourital 01 Industrial and Engineering Chemistry: On the appearance of hlr. Bailey’s paper on “Accuracy in Sampling Coal”* I gave it considerable study. By the time I had arrived a t my conclusions in regard to it, interest, judging by cessation of correspondence, had waned. I therefore hesitated to publish my views. The recrudescence of Mr. Bailey’s paper in that of Mr. F. B. P ~ r t e r however, ,~ seems to warrant calling attention to some errors in the former. Mr. Bailey began his paper by comparing some results obtained in the A. D. Little Laboratory with the results tabulated on pp. 281284, Professional Paper No. 48, U. S Geological Survey. This tabulation led Mr. Bailey into a n error which may best be explained by a concrete instance. 4833 Rs. of an Iowa coal were sampled and analyzed. This lot showed 20.70 per cent. ash, 10986 Rs. from the same carload showed 16.04per cent. ash, 12000 Rs. showed an ash of 14.46per cent. Of course, the proper way to obtain an average analysis is to multiply each analysis by the weight of the coal which i t represents, add the products and divide by the total weight of coal. Calculated in this manner the coal in question showed 16.17per cent. ash. The carload from which these portions were taken was sampled as a whole and showed 15.22 per cent. ash. The best value available for this coal is an average of the tw figures, 15.69per cent., and the error due to sampling is 0.47 per cent. In the averaging of the Geological Survey the figures were averaged, as much weight being allowed to the analysis representing 4833 Bs. as to the analysis representing the carload. Calculated in this way the average was 16.60 per cent., and the difference between that figure and 20.70 per cent., or 4.10per cent., Mr. Bailey considered error. Recalculating this tabulation the maximum error is 1.16per cent. and only 5 analyses out of 34 show an error greater than 0.7 per cent. It is probable that these higher errors would disappear if we could trace all the coal involved, but as it stands the maximum error is much less than that of the commercial laboratory as might reasonably be expected from the favorable conditions under which the government chemists worked. Whereas Mr. Bailey thought the government error was more than twice as large as the other. Mr. Bailey closes his comparison of these two sets of figures with the following statement: “As this data follows the probability curve very closely, the formula for probable and possible error can be safely applied in determining the reliability of different methods of sampling.” Even if Mr. Bailey had substantiated his formulae by correct figures there would be good reason to be suspicious of them on a Comfit. rend.. 150,540 and 151, 78. (1909). 3 Ibzd I, 148 (1910). 1
THISJOURNAL, 1, 161
